EFF_D_SVM: a robust multi-type brain tumor classification system.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Brain tumors pose a significant threat to human health, necessitating accurate and efficient diagnostic methods.
• Automated brain tumor diagnosis systems using Magnetic Resonance Imaging (MRI) can aid clinicians and reduce workload.
• The scarcity of brain tumor data makes leveraging pre-trained Convolutional Neural Network (CNN) models a practical strategy for classification.

Purpose of the Study:

• To propose and evaluate a novel brain tumor classification system, EFF_D_SVM, designed to improve diagnostic accuracy.
• To introduce a new feature extraction module, EFF_D, integrated with the EfficientNetB0 model.
• To validate the system's performance against existing state-of-the-art models.

Main Methods:

• Developed the EFF_D_SVM system by modifying the EfficientNetB0 model with a new feature extraction module (EFF_D).
• Fine-tuned the EFF_D module using Softmax for feature extraction from brain tumor MRI images.
• Classified extracted features using Support Vector Machine (SVM) and employed Grad-CAM for visualization.

Main Results:

• The EFF_D_SVM system demonstrated superior performance compared to other state-of-the-art models across key evaluation metrics.
• Comparative experiments and cross-validation confirmed the proposed model's effectiveness and robustness.
• Grad-CAM visualization provided insights into the features extracted from brain tumor images.

Conclusions:

• The proposed EFF_D_SVM system offers a highly effective and robust solution for automated brain tumor classification from MRI data.
• The novel feature extraction module and SVM classification contribute to improved diagnostic accuracy.
• This approach holds significant potential for clinical application in brain tumor diagnosis.